Well, an accelerometer is a promising start, but you need to be very cautious. Resonance patterns in loudspeaker enclosures are all over the place, implying you must measure at hundreds of points in systematic manner to get the picture and literally map the entire surface
The usual Stereophile single point accelerometer measurements do not tell us much.
Good luck,
Eelco
The usual Stereophile single point accelerometer measurements do not tell us much.
Good luck,
Eelco
What is the relationship between the acceleration you measure at a point on a panel and what you hear as cabinet radiation at the listening position?
There are a number of significant problems with measuring acceleration at a point on a panel and trying to relate it to what is heard as cabinet noise. For example:
- the magnitude of the vibration varies strongly over the panel. There are areas of strong vibration and areas of no vibration and they change with frequency.
- the phase of the vibration varies strongly over the panel. Sound radiated by different areas can be both of a different phase and a significant proportion of a wavelength apart causing substantial cancellation and reinforcement in the far field.
- Energy transferred from cabinet motion to sound waves in one area can be transferred back from the sound waves into the cabinet at another.
If you want to get a handle on the cabinet sound level at the listening position you will need to measure the vibration over the whole loudspeaker making sure you get the relative phase as well as amplitude. Then sum all the sound sources over the speaker to get the sound at the listening position and other locations of interest. Of course, the sound radiation is also affected by the room (e.g. a strongly vibrating rear panel in an anechoic chamber will be less of a problem than in a room close to the front wall) so that probably ought to be included when summing the sound sources over the cabinet surface.
Why?
At frequencies below the resonances, typically a few hundred Hz, where mass*acceleration is small compared to stiffness*deflection you believe a good approach to achieving the same level of panel deflection is to use heavy limp panels rather than light stiff panels?
No it is not. Below resonance stiffness dominates and a limp heavy panel will generate relatively large panel deflections under the driver forces whereas a light stiffer panel will deflect less. At high frequency things are different because mass*acceleration will increase with frequency squared whereas stiffness*deflection is independent of frequency.
If you want to dispute the above can you please provide reasoning rather than simply making statements.
In reference to Andy's statements above, and from one of my prior posts on another thread at this forum:
For basic panel vibration concepts:
Stiffness controlled (Low frequencies)
Fundamental resonance frequency region, damping impacted
Mass controlled (Mid frequencies)
Coincidence frequency region, damping impacted
Damping controlled (High frequencies)
http://personal.inet.fi/koti/juhladude/soundproofing.html
For basic panel vibration concepts:
Stiffness controlled (Low frequencies)
Fundamental resonance frequency region, damping impacted
Mass controlled (Mid frequencies)
Coincidence frequency region, damping impacted
Damping controlled (High frequencies)
http://personal.inet.fi/koti/juhladude/soundproofing.html
Last edited:
No it is not.
Below resonance stiffness dominates and
a limp heavy panel will generate relatively large panel deflections under the driver forces whereas a light stiffer panel will deflect less.
At high frequency things are different because mass*acceleration will increase with frequency squared whereas stiffness*deflection is independent of frequency.
If you want to dispute the above can you please provide reasoning rather than simply making statements.
No, it is not not.
Ad
: there may be different resonant frequencies/modes, they are all caused by a mass coupled to a spring. So, where does stiffness drop out of the equation. Or mass?Ad
: it stands to reason that something heavy will deflect less than something light under the same application of force. Don't forget that the pressure differentials between inside and outside of the enclosure are not the main cause of enclosure vibrations. Reaction forces are. Stiffness really doesn't come into play (except for providing the springiness required to generate resonances).ad
: The F in Newton's second law stands for force, not frequency. But I may have misunderstood your point.
When complex problems like this get grossly oversimplified, like this has, then conflicts are inevitable. It is true that the major force excitation is the driver itself and that the larger the mass of the panels the lower the acceleration. But it is also true that the stiffness controls the displacement of the force at LFs and that the panels have different regions, like stiffness controlled, coincidence, etc., but these definitions are strictly applied only to very large panels where the waves actually propagate in them. It is dangerous to try and apply them to finite panels and resonant structures.
The structure resonates, that much we know, but beyond that things get more complex still. Just finding the displacements of the panels does not yield the sound radiated. This is a much more complex problem. It is very possible to have large panel motions and near zero net sound radiation - it may not be common, but it is certainly possible.
To make any sense of this one must look at it statistically. Clearly a lower average motion will yield a lower average sound radiation. But statistically it is the damping that will dominate the problem. The higher the damping the lower the average motion. Damping is less effective on massive objects and hence a lighter stiffer structure will tend to have less motion, excluding the fact that the force will be greater on the lighter enclosure. It seems to me and has always seemed this way, that damping is key and that a highly damped front panel where the drivers are mounted is crucial. Lighter, stiffer and well damped materials seem to work best in lowering the net average motion, but how the enclosure is constructed - with bracing, damped braces, etc. is probably more important than the actual material.
The structure resonates, that much we know, but beyond that things get more complex still. Just finding the displacements of the panels does not yield the sound radiated. This is a much more complex problem. It is very possible to have large panel motions and near zero net sound radiation - it may not be common, but it is certainly possible.
To make any sense of this one must look at it statistically. Clearly a lower average motion will yield a lower average sound radiation. But statistically it is the damping that will dominate the problem. The higher the damping the lower the average motion. Damping is less effective on massive objects and hence a lighter stiffer structure will tend to have less motion, excluding the fact that the force will be greater on the lighter enclosure. It seems to me and has always seemed this way, that damping is key and that a highly damped front panel where the drivers are mounted is crucial. Lighter, stiffer and well damped materials seem to work best in lowering the net average motion, but how the enclosure is constructed - with bracing, damped braces, etc. is probably more important than the actual material.
Last edited:
I am contemplating a new speaker build but will probably find it hard to swing the WAF in my favor in the foreseeable future. This does mean that I have a lot of time to research all options and design philosophies.
I'm not quite a standard layperson as I have a degree in composite engineering and my dissertation involved resonance control of composite panels. However this was 15yrs ago and my calculus isn't quite as sharp as it used to be.
When building my current speakers, about 5yrs ago, I broke one of my main design mantras. "don't do anything just because everyone else is doing it that way unless you fully understand why it's done that way and agree" . Everyone was using mdf so that is what I used. Mistake. Don't get me wrong, my speakers sound very good but I think the material used is sucking a bit of life from the music.
I have to agree with Dave (planet 10) about keeping all fundamental frequencies out of the passband by pushing them either higher or lower than the xo for that enclosure. Also that excessive use of damping deadens and kills the sound.
A material that was briefly mentioned earlier in this thread was corian. Corian is acrylic with a filler made of finely ground ceramic/rock. Corian is primarily used for kitchen counter tops and the filler is added for aesthetic reasons and to give the counter a tough scratch resistant surface. Both of these attributes could be desirable. The addition of the filler also dramatically changes the mechanical properties of this material compared to straight acrylic. It is much denser, has higher damping ratio and slightly more rigid. The exact properties will depend on the filler used as they are selected for colour primarily. I doubt that corian is an ideal solution for cabinets covering the lower frequencies but might be suitable for upper midrange/treble enclosures.
Straight acrylic on the other hand should be an excellent material. Although it is denser than ply or mdf it is also much more rigid and has a higher specific modulus(modulus of elasticity divided by density). It also has good self damping but isn't over-damped. It is much cheaper and easier to source than corian and is available in many thicknesses from 1mm to over 100mm(though the thickest gauges are more expensive and difficult to source). An acrylic panel of the same size and thickness as a ply panel will have a higher resonate frequency. Build a base bin out of 20mm acrylic with lots of internal braces and you'll still have a cabinet with primary panel resonances well out of the bass region and enough mass to satisfy most more mass adherents.
To join panels together a solvent is used that dissolves the surfaces in the joint then evaporates leaving a single homogeneous solid as if it were hewn from a solid block.
For midrange cabinets many thinner baffle/braces could be used say 6mm thickness spaced every 25mm or so to give an incredibly rigid structure. The multiple braces would also help to break up the rear wave and allow precise location of wadding where it is needed most, in the centre.
Theoretically acrylic is an excellent material but unfortunately I am not currently in a situation that will allow me to test to see if it actually is.
Bose use to make display versions of some of their speakers out of acrylic rather than the chipboard used for production models. They sounded god awful but a hell of a lot less awful than the chipboard ones.
On a completely different line of thought for those of you using plywood.
This comes straight from my research with composite materials which is all about aligning the fibres with the greatest efficiency.
Have any of you tried cutting the plywood on the bias. This would mean cutting the panels so the lamina are at 45degrees/45degrees rather than 90/90. This would of course result in a lot of wasted triangular offcuts but should increase panel rigidity by 25-40%.
Keep up the lively debate and keep teaching me new stuff.
Niffy
I'm not quite a standard layperson as I have a degree in composite engineering and my dissertation involved resonance control of composite panels. However this was 15yrs ago and my calculus isn't quite as sharp as it used to be.
When building my current speakers, about 5yrs ago, I broke one of my main design mantras. "don't do anything just because everyone else is doing it that way unless you fully understand why it's done that way and agree" . Everyone was using mdf so that is what I used. Mistake. Don't get me wrong, my speakers sound very good but I think the material used is sucking a bit of life from the music.
I have to agree with Dave (planet 10) about keeping all fundamental frequencies out of the passband by pushing them either higher or lower than the xo for that enclosure. Also that excessive use of damping deadens and kills the sound.
A material that was briefly mentioned earlier in this thread was corian. Corian is acrylic with a filler made of finely ground ceramic/rock. Corian is primarily used for kitchen counter tops and the filler is added for aesthetic reasons and to give the counter a tough scratch resistant surface. Both of these attributes could be desirable. The addition of the filler also dramatically changes the mechanical properties of this material compared to straight acrylic. It is much denser, has higher damping ratio and slightly more rigid. The exact properties will depend on the filler used as they are selected for colour primarily. I doubt that corian is an ideal solution for cabinets covering the lower frequencies but might be suitable for upper midrange/treble enclosures.
Straight acrylic on the other hand should be an excellent material. Although it is denser than ply or mdf it is also much more rigid and has a higher specific modulus(modulus of elasticity divided by density). It also has good self damping but isn't over-damped. It is much cheaper and easier to source than corian and is available in many thicknesses from 1mm to over 100mm(though the thickest gauges are more expensive and difficult to source). An acrylic panel of the same size and thickness as a ply panel will have a higher resonate frequency. Build a base bin out of 20mm acrylic with lots of internal braces and you'll still have a cabinet with primary panel resonances well out of the bass region and enough mass to satisfy most more mass adherents.
To join panels together a solvent is used that dissolves the surfaces in the joint then evaporates leaving a single homogeneous solid as if it were hewn from a solid block.
For midrange cabinets many thinner baffle/braces could be used say 6mm thickness spaced every 25mm or so to give an incredibly rigid structure. The multiple braces would also help to break up the rear wave and allow precise location of wadding where it is needed most, in the centre.
Theoretically acrylic is an excellent material but unfortunately I am not currently in a situation that will allow me to test to see if it actually is.
Bose use to make display versions of some of their speakers out of acrylic rather than the chipboard used for production models. They sounded god awful but a hell of a lot less awful than the chipboard ones.
On a completely different line of thought for those of you using plywood.
This comes straight from my research with composite materials which is all about aligning the fibres with the greatest efficiency.
Have any of you tried cutting the plywood on the bias. This would mean cutting the panels so the lamina are at 45degrees/45degrees rather than 90/90. This would of course result in a lot of wasted triangular offcuts but should increase panel rigidity by 25-40%.
Keep up the lively debate and keep teaching me new stuff.
Niffy
Damping factors for some materials:
damping factor values : damping factors
more damping factor values : damping factors
damping factor values : damping factors
more damping factor values : damping factors
Are you implying that sound radiation from the enclosure enhances the sound? Makes it livelier. Lower damping of any component in a loudspeaker is going to add resonant coloration. I always thought that was a bad thing.
Not quite.
If the cabinet had very low damping it would ring like a bell if excited but the frequency range that could excite it would also be narrow, a tall thin spike. As long as no excitable frequencies are introduced no problem.
Higher damping results in a broader shallower spike meaning that a wider range of frequencies can excite the cabinet. It is therefore more difficult to keep the frequencies that can excite the cabinet out of the passband especially as the addition of damping material inevitably also moves the resonant frequency lower. So lightly damped is the way to go.
Niffy
I can't remember there name but there is a speaker company that uses solid wood (and the designs looks like a double bass) that 'tune' their wooden cabinets to enhance the sound as a musical instrument would. Seems like the polar opposite approach of most to make the least radiant cabinets by what ever means.
Inreresting charts indeed.
Seems that Simple Stoopid Particle board, as Beloved by Many speaker makers in the 70's/80's And IKEA has significant merit.. beyond it's lowly status.
Better seemingly than a Lot of Highly Touted (pricey) Materials.. certainly so in the pedestrian /noncomposite types.
Emperors' Clothes is seemingly a Perennial element in Audio Weenie Land
Seems that Simple Stoopid Particle board, as Beloved by Many speaker makers in the 70's/80's And IKEA has significant merit.. beyond it's lowly status.
Better seemingly than a Lot of Highly Touted (pricey) Materials.. certainly so in the pedestrian /noncomposite types.
Emperors' Clothes is seemingly a Perennial element in Audio Weenie Land